I started a post at iMaterilist 2020: No points/prizes are killing research competitions where I try to explain why I think these changes will adversely affect to these research competitions.
\n\nUnless you have plenty of resources or this is your research field, I don't see how these competitions can encourage/motivate competitors and CV experts to solve these problems…
\n\nLast year iMet, iMaterialist, Kuzushiji Recognition, Recursion Cellular, Human Atlas, all the Google comps! I think those research competitions were great, a lot of people competed and for that reason the solutions were great too, as the rewards :)
\n\nI think the reason why kaggle works is because a lot of people compete that's the point imo. Motivate people to join, get interested in the subject and compete… and obviously have fun! An easy way to encourage/motivate top kagglers who can provide such solutions (and will do in many cases) is to offer great prizes (for example Kuzushiji Recognition Competition).
\n\n> I'm wondering if these changes comes due to last year controversial results at this competition...
\n\nAnyway, I hope people participates here, so these competitions will keep alive.
", "messages": [ { "id": "788401", "postDate": "03/27/2020 16:34:27", "content": "I started a post at iMaterilist 2020: No points/prizes are killing research competitions where I try to explain why I think these changes will adversely affect to these research competitions.
\n\nUnless you have plenty of resources or this is your research field, I don't see how these competitions can encourage/motivate competitors and CV experts to solve these problems…
\n\nLast year iMet, iMaterialist, Kuzushiji Recognition, Recursion Cellular, Human Atlas, all the Google comps! I think those research competitions were great, a lot of people competed and for that reason the solutions were great too, as the rewards :)
\n\nI think the reason why kaggle works is because a lot of people compete that's the point imo. Motivate people to join, get interested in the subject and compete… and obviously have fun! An easy way to encourage/motivate top kagglers who can provide such solutions (and will do in many cases) is to offer great prizes (for example Kuzushiji Recognition Competition).
\n\n> I'm wondering if these changes comes due to last year controversial results at this competition...
\n\nAnyway, I hope people participates here, so these competitions will keep alive.
", "rawMarkdown": "I started a post at iMaterilist 2020: [No points/prizes are killing research competitions](https://www.kaggle.com/c/imaterialist-fashion-2020-fgvc7/discussion/138107) where I try to explain why I think these changes will adversely affect to these research competitions.\n\nUnless you have plenty of resources or this is your research field, I don't see how these competitions can encourage/motivate competitors and CV experts to solve these problems…\n\nLast year iMet, iMaterialist, Kuzushiji Recognition, Recursion Cellular, Human Atlas, all the Google comps! I think those research competitions were great, a lot of people competed and for that reason the solutions were great too, as the rewards :)\n\nI think the reason why kaggle works is because a lot of people compete that's the point imo. Motivate people to join, get interested in the subject and compete… and obviously have fun! An easy way to encourage/motivate top kagglers who can provide such solutions (and will do in many cases) is to offer great prizes (for example Kuzushiji Recognition Competition).\n\n> I'm wondering if these changes comes due to **last year controversial results** at this competition...\n\nAnyway, I hope people participates here, so these competitions will keep alive.", "votes": null }, { "id": "788523", "postDate": "03/27/2020 18:37:13", "content": "Thanks for your post!
", "rawMarkdown": "Thanks for your post!", "votes": null }, { "id": "788697", "postDate": "03/27/2020 22:36:01", "content": "Sad, I was waiting this competition. But it's understandable decision. Labels are available online so the catastrophe can happen like last year.
", "rawMarkdown": "Sad, I was waiting this competition. But it's understandable decision. Labels are available online so the catastrophe can happen like last year.", "votes": null }, { "id": "788743", "postDate": "03/28/2020 00:35:54", "content": "Yes, I guess that is one of the reasons. Sad, but I respect it and I will play with the data :)
", "rawMarkdown": "Yes, I guess that is one of the reasons. Sad, but I respect it and I will play with the data :)", "votes": null }, { "id": "788746", "postDate": "03/28/2020 00:43:29", "content": "I feel the same as you.
\n\nThe last year's iMet was my first solo CV competition at Kaggle, I learned many things in it.\nSince I made a mistake and lost my place there, I have been waiting for a long long loooooong time for challenging it again!
\n\nI understand medals or prizes can cause such a terrible dark deed like X5, but no medal nor prize cannot bring out participants' amazing and incredible efforts such as the last year's 1st solution.
\n\ni will participate in this year's iMet, but not so passionately as last year...
\n\nI'm so sad that I cannot compete with kagglers who try their utmost.\n
\nThanks.
Thanks for your post @jesucristo. I feel the same as you.
", "rawMarkdown": "Thanks for your post @jesucristo. I feel the same as you.", "votes": null }, { "id": "809334", "postDate": "04/16/2020 04:43:38", "content": "Maybe @drn01z3 can answer what was the intention last year to scrape the labels? Looks like you drastically hurt the FGVC series with your actions.
", "rawMarkdown": "Maybe @drn01z3 can answer what was the intention last year to scrape the labels? Looks like you drastically hurt the FGVC series with your actions.", "votes": null }, { "id": "1393751", "postDate": "07/19/2021 22:02:40", "content": "There is a lot of truth in this post.
", "rawMarkdown": "There is a lot of truth in this post.", "votes": null } ], "comments": [ { "id": 1393751, "author_name": "georgidragomanov", "author_url": "", "post_date": "07/19/2021 22:02:40", "content": "There is a lot of truth in this post.
", "votes": null, "replies": [] }, { "id": 788523, "author_name": "codingcliff", "author_url": "", "post_date": "03/27/2020 18:37:13", "content": "Thanks for your post!
", "votes": null, "replies": [] }, { "id": 788697, "author_name": "bamps53", "author_url": "", "post_date": "03/27/2020 22:36:01", "content": "Sad, I was waiting this competition. But it's understandable decision. Labels are available online so the catastrophe can happen like last year.
", "votes": null, "replies": [ { "id": 788743, "author_name": "jesucristo", "author_url": "", "post_date": "03/28/2020 00:35:54", "content": "Yes, I guess that is one of the reasons. Sad, but I respect it and I will play with the data :)
", "votes": null, "replies": [] } ] }, { "id": 788746, "author_name": "ttahara", "author_url": "", "post_date": "03/28/2020 00:43:29", "content": "I feel the same as you.
\n\nThe last year's iMet was my first solo CV competition at Kaggle, I learned many things in it.\nSince I made a mistake and lost my place there, I have been waiting for a long long loooooong time for challenging it again!
\n\nI understand medals or prizes can cause such a terrible dark deed like X5, but no medal nor prize cannot bring out participants' amazing and incredible efforts such as the last year's 1st solution.
\n\ni will participate in this year's iMet, but not so passionately as last year...
\n\nI'm so sad that I cannot compete with kagglers who try their utmost.\n
\nThanks.
Thanks for your post @jesucristo. I feel the same as you.
", "votes": null, "replies": [] }, { "id": 809334, "author_name": "christofhenkel", "author_url": "", "post_date": "04/16/2020 04:43:38", "content": "Maybe @drn01z3 can answer what was the intention last year to scrape the labels? Looks like you drastically hurt the FGVC series with your actions.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "788401": "I started a post at iMaterilist 2020: [No points/prizes are killing research competitions](https://www.kaggle.com/c/imaterialist-fashion-2020-fgvc7/discussion/138107) where I try to explain why I think these changes will adversely affect to these research competitions.\n\nUnless you have plenty of resources or this is your research field, I don't see how these competitions can encourage/motivate competitors and CV experts to solve these problems…\n\nLast year iMet, iMaterialist, Kuzushiji Recognition, Recursion Cellular, Human Atlas, all the Google comps! I think those research competitions were great, a lot of people competed and for that reason the solutions were great too, as the rewards :)\n\nI think the reason why kaggle works is because a lot of people compete that's the point imo. Motivate people to join, get interested in the subject and compete… and obviously have fun! An easy way to encourage/motivate top kagglers who can provide such solutions (and will do in many cases) is to offer great prizes (for example Kuzushiji Recognition Competition).\n\n> I'm wondering if these changes comes due to **last year controversial results** at this competition...\n\nAnyway, I hope people participates here, so these competitions will keep alive.", "788523": "Thanks for your post!", "788697": "Sad, I was waiting this competition. But it's understandable decision. Labels are available online so the catastrophe can happen like last year.", "788743": "Yes, I guess that is one of the reasons. Sad, but I respect it and I will play with the data :)", "788746": "I feel the same as you.\n\n\nThe last year's iMet was my first solo CV competition at Kaggle, I learned many things in it.\nSince I made a mistake and lost my place there, I have been waiting for a long long loooooong time for challenging it again! \n\n\nI understand medals or prizes can cause such a terrible dark deed like _X5_, but no medal nor prize cannot bring out participants' amazing and incredible efforts such as [the last year's 1st solution](https://www.kaggle.com/c/imet-2019-fgvc6/discussion/94687).\n\n\ni will participate in this year's iMet, but not so passionately as last year...\n\nI'm so sad that **I cannot compete with kagglers who try their utmost**.\n